GudDesk
Docs
Knowledge Base

Knowledge Base

Create a help center with articles that customers and AI agents can search.

The GudDesk knowledge base lets you create help articles that serve two purposes: customers can browse them for self-service, and your AI agents use them as their source of truth when answering questions.

Creating Articles

  1. Navigate to Dashboard > Knowledge Base
  2. Click New Article
  3. Write your article using the rich text editor (supports Markdown)
  4. Assign it to a collection (category)
  5. Set the status to Published when ready

Article Structure

Good help articles follow a consistent structure:

  • Title — Clear, searchable (e.g., "How to Reset Your Password")
  • Description — A brief summary shown in search results
  • Body — Step-by-step instructions with screenshots when helpful
  • Collection — The category this article belongs to

Collections

Collections organize your articles into browsable categories:

  • Getting Started — Onboarding guides for new users
  • Account & Billing — Password resets, plan changes, invoices
  • Features — How to use specific product features
  • Troubleshooting — Common issues and fixes

Create collections in Settings > Knowledge Base > Collections.

Public Help Center

Your knowledge base is automatically available as a public help center at:

https://your-domain.com/help

Customers can browse collections, search articles, and find answers without contacting support.

AI Agent Integration

This is where the knowledge base becomes powerful. When a customer asks a question through the chat widget:

  1. The AI agent searches your knowledge base for relevant articles
  2. If it finds a match, it summarizes the answer and links to the full article
  3. If the customer confirms the answer helped, the conversation is auto-resolved
  4. If not, the agent escalates to a human

The more comprehensive your knowledge base, the more conversations your AI agents can resolve automatically.

The knowledge base includes full-text search powered by PostgreSQL. Articles are indexed by title, description, and body content. Search results are ranked by relevance.

Analytics

Track how your knowledge base performs:

  • Article views — Which articles are most popular
  • Search queries — What customers are searching for (use this to identify gaps)
  • Resolution rate — How often an article view prevents a support conversation

Next Steps